size mismatch for rpn.rpn_reg_layer

[Niceday99]

Hi guys, when I run eval_rcnn.py with the rcnn model trained by myself, following error occurs: And it's ok if I use the model provided by author, anyone can help me?

[adioshun]

RuntimeError: Error(s) in loading state_dict for PointRCNN:
	size mismatch for rpn.rpn_reg_layer.2.conv.weight: copying a param with shape torch.Size([52, 128, 1]) from checkpoint, the shape in current model is torch.Size([76, 128, 1]).
	size mismatch for rpn.rpn_reg_layer.2.conv.bias: copying a param with shape torch.Size([52]) from checkpoint, the shape in current model is torch.Size([76]).

same problem. Did you solve it?

[saurabverma]

LOC_XZ_FINE = True >>> Size = 76 LOC_XZ_FINE = False >>> Size = 52

[jingyi-zhang]

set 'LOC_XZ_FINE:False' in default.yaml